ofthe
ROYAL ASTRONOMICAL SOCIETY
MNRAS Advance Access published February 4, 2013
MNRAS (2013)
doi: 1 0. 1 093/mnras/ sts692
Unification of X-ray winds in Seyfert galaxies: from ultra-fast outflows
to warm absorbers
F. Tombesi, 1,2 * M. Cappi , 3 J. N. Reeves , 4 R. S. Nemmen , 1 V. Braito , 5 M. Gaspari 6 and
C. S. Reynolds 2
} NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA
2 Department of Astronomy, University of Maryland, College Park, MD 20742, USA
3 INAF-IASF Bologna, Via Gobetti 101, 1-40129 Bologna, Italy
A Astrophysics Group, School of Physical and Geographical Sciences, Keele University, Keele, Staffordshire ST5 5BG, UK
5 INAF - Osservatorio Astronomico di Brera, via. E. Bianchi 46, 1-23807 Merate, Italy
6 Max Planck Institute for Astrophysics, Karl-Schwarzschild-Strasse 1, D -857 41 Garching, Germany
Accepted 2012 December 19. Received 2012 December 19; in original form 2012 November 7
ABSTRACT
The existence of ionized X-ray absorbing layers of gas along the line of sight to the nuclei
of Seyfert galaxies is a well established observational fact. This material is systematically
outflowing and shows a large range in parameters. However, its actual nature and dynamics
are still not clear. In order to gain insights into these important issues we performed a literature
search for papers reporting the parameters of the soft X-ray warm absorbers (WAs) in 35 type
1 Seyferts and compared their properties to those of the ultra-fast outflows (UFOs) detected in
the same sample. The fraction of sources with WAs is >60 per cent, consistent with previous
studies. The fraction of sources with UFOs is >34 per cent, >67 per cent of which also
show WAs. The large dynamic range obtained when considering all the absorbers together,
spanning several orders of magnitude in ionization, column, velocity and distance allows us,
for the first time, to investigate general relations among them. In particular, we find significant
correlations indicating that the closer the absorber is to the central black hole, the higher the
ionization, column, outflow velocity and consequently the mechanical power. In all the cases,
the absorbers continuously populate the whole parameter space, with the WAs and the UFOs
lying always at the two ends of the distribution. These evidence strongly suggest that these
absorbers, often considered of different types, could actually represent parts of a single large-
scale stratified outflow observed at different locations from the black hole. The UFOs are likely
launched from the inner accretion disc and the WAs at larger distances, such as the outer disc
and/or torus. We argue that the observed parameters and correlations are, to date, consistent
with both radiation pressure through Compton scattering and magnetohydrodynamic processes
contributing to the outflow acceleration, the latter playing a major role. Most of the absorbers,
especially the UFOs, show a sufficiently high mechanical power (at least ~0.5 per cent of the
bolometric luminosity) to provide a significant contribution to active galactic nuclei (AGN)
feedback and thus to the evolution of the host galaxy. In this regard, we find possible evidence
for the interaction of the AGN wind with the surrounding environment on large scales.
Key words: accretion, accretion discs -black hole physics -galaxies: active -galaxies:
Seyfert -X-rays: galaxies.
1 INTRODUCTION
^E-mail: ftombesi@astro.umd.edu
The presence of ionized material along the line of sight to the
nuclei of Seyfert galaxies has been known for a long time (e.g.
Halpern 1984). The material observable in absorption in the soft
X-rays has been referred to as a warm absorber (WA). The limited
©2013 The Authors
Published by Oxford University Press on behalf of the Royal Astronomical Society
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
2 F. Tombesi et al.
energy resolution of previous X-ray instruments allowed essentially
only the detection of broad absorption edges. In fact, observations
with Advanced Satellite for Cosmology and Astrophysics detected
O vn and O vm absorption edges and established that WAs are a
common feature of active galactic nuclei (AGNs), being present
in ~50 per cent of Seyfert galaxies (Reynolds 1997; George et al.
1998). Then, the advent of the unprecedented spectral resolution of
the gratings on board Chandra and XMM-Newton allowed, for the
first time, the detection of discrete soft X-ray resonance absorption
and emission lines. The resulting picture of the WA is that of an
outflow exhibiting multiple narrow absorption lines corresponding
to different ionization states (Kaastra et al. 2000; Kaspi et al. 2000;
Blustin et al. 2005; McKernan, Yaqoob & Reynolds 2007). The
values of the ionization parameter are typically in the range log £ ~
0-2 erg s _1 cm and the column densities are between /V H ~ 10 20
and 10 22 cm -2 . The absorption lines are systematically blue-shifted,
indicating outflow velocities of the WAs in the range v out ~ 100-
1000 km s -1 . There are still significant uncertainties on the exact
location of this material, which ranges from ~pc up to ~kpc scales,
and it has been suggested that it might originate outside of the inner
disc, probably at locations comparable with the obscuring torus
(e.g. Krolik & Kriss 2001; Blustin et al. 2005). Depending on the
actual filling and covering factors, the mass outflow rate from the
WAs can be significant but, given their relatively low velocities,
their kinetic power is rather low when compared to the bolometric
luminosity (e.g. Blustin et al. 2005; McKernan et al 2007). However,
a recent detailed study by Crenshaw & Kraemer (2012) found that
the integrated power of the WAs in some Seyferts can actually
reach the level of ~0. 1-0.5 per cent of the bolometric luminosity,
the minimum required by numerical simulations for AGN feedback
to exert a significant impact on the host galaxy (e.g. Hopkins & Elvis
2010; Gaspari et al. 201 la,b; Gaspari, Brighenti & Temi 2012b).
Besides WAs, highly blueshifted Fe K-shell absorption lines at
E >1 keV have been detected in more recent years in the X-
ray spectra of several AGNs (Chartas et al. 2002, 2009; Chartas,
Brandt & Gallagher 2003; Pounds et al. 2003; Dadina et al. 2005;
Markowitz, Reeves & Braito 2006; Braito et al. 2007; Cappi et al.
2009; Reeves et al. 2009; Giustini et al. 2011; Gofford et al. 2011;
Lobban et al. 2011; Dauser et al. 2012). In particular, a uniform and
systematic search for blueshifted Fe K absorption lines in a sample
of 42 local (z < 0. 1) Seyferts observed with XMM-Newton was per-
formed by Tombesi et al. (2010a, hereafter Paper I). This allowed
the authors to assess their global significance and derive a detection
fraction of >40 per cent. In order to mark an initial and somewhat
arbitrary distinction with the classical WAs, in paper I we defined
ultra-fast outflows (UFOs) as those highly ionized Fe K absorbers
with a blueshifted velocity >10 000 km s _1 . Subsequently, Tombesi
et al. (201 la, hereafter Paper II) performed a photo-ionization mod-
elling of these UFOs and derived the distribution of their main
physical parameters. The outflow velocity is mildly relativistic, in
the range ~0.03-0.3c (~10 000-100 000 km s -1 ), with mean value
of ~0.1c. The ionization is very high, in the range log£ ~ 3-
6 erg s -1 cm, with a mean value of ~4.2 erg s -1 cm. The column
densities are also large, in the interval /V H ~ 10 22 -10 24 cm -2 . These
findings are important because they suggest the presence of mas-
sive and highly ionized absorbers outflowing with mildly relativistic
velocities from the nuclei of these Seyfert galaxies.
In a following paper, Tombesi et al. (2012a, hereafter Paper III)
quantified that they are observable at locations of sub-parsec scales
from the central black hole. Their mass outflow rate was constrained
between ~0.01 and 1 Mq yr -1 and their kinetic power was found
in the range \ogE K — 42-45 erg s -1 . Thus, the UFOs are possi-
bly directly identifiable, at least qualitatively, with accretion disc
winds/outflows (King & Pounds 2003; Proga & Kallman 2004;
Ohsuga et al. 2009; Sim et al. 2010; Fukumura et al. 2010) or the
base of a possible weak jet (e.g. Ghisellini, Haardt & Matt 2004).
In particular, the kinetic power of these UFOs is systematically
higher than the minimum required by simulations of feedback in-
duced by winds/outflows (e.g. Hopkins & Elvis 2010; Gaspari et al.
2011a,b, 2012b). Therefore, in the long term, they could be able
to significantly influence the bulge evolution, star formation, super-
massive black hole growth and contribute to the establishment of
the observed black hole-host galaxy relations, such as the M^-cr
(Ferrarese & Merritt 2000; King 2010a; Ostriker et al. 2010; Gas-
pari et al. 201 la,b; Gaspari, Ruszkowski & Sharma 2012a; Gaspari
et al. 2012b).
It is important to note that Tombesi et al. (2010b, 2011b, 2012b)
detected the presence of UFOs also in a small sample of (three
out of five) radio-loud AGNs observed with Suzaku. Finally, we
also note that similar results regarding the UFOs have been ob-
tained independently by Gofford et al. (2012) who performed a
uniform and systematic broad-band spectral analysis of a large
sample of AGNs observed with Suzaku , confirming their overall
incidence and characteristics. Moreover, evidence for similar out-
flows are emerging also in stellar-mass black holes (e.g. King et al.
2012 ).
Despite these significant observational developments, the origin
and acceleration mechanism(s) of the ionized absorbers in AGNs
are still debated. In particular, radiation and magnetohydrodynamic
(MHD) wind models have been developed and scenarios in which
they are considered as intrinsically distinct or as different manifes-
tations of the same phenomenon have been suggested (e.g. Konigl
& Kartje 1994; Elvis 2000; Krolik & Kriss 2001; Blustin et al.
2005; Krongold et al. 2007; Fukumura et al. 2010; Kazanas et al.
2012; Reynolds 2012). Given the relevance of these outflows for
the physics and energetics of AGNs and their potential significant
contribution to feedback, it is imperative to investigate them in
detail.
In order to test these hypothesis, in this paper we will perform a
detailed observational comparison of the WAs and UFOs, checking
for correlations and discussing the possible unification of these
absorbers in a single, photo-ionized and stratified outflow. We focus
on the sample of Seyfert galaxies described in Paper I and will use
WA parameters collected from the literature and those of the UFOs
derived in Papers II and III, respectively.
2 WARM ABSORBERS SELECTION
We performed a literature search for papers reporting the analysis
of the soft X-ray WAs in the 35 type 1 Seyferts of the sample dis-
cussed in Paper I. This was defined selecting all the Seyfert galaxies
from the RXTE All-Sky Slew Survey Catalogue (Revnivtsev et al.
2004) and cross-correlating them with the XMM-Newton Accepted
Targets Catalogue (as of October 2008). After applying the stan-
dard filtering processes, we obtained a total of 42 objects with 101
good XMM-Newton observations and, more specifically, 35 clas-
sified as type 1 and seven as type 2 with 87 and 14 observations,
respectively.
In the literature search, we selected only the WA results de-
rived from the high-energy resolution gratings on board Chandra
and XMM-Newton because they allow us to constrain the outflow
velocity, which is crucial to derive the mass outflow rate and the
kinetic power. We limited our search only to the type 1 sources be-
cause, in accordance with the unification model, the spectra of the
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies 3
seven type 2s are affected by significant neutral absorption (A/ H >
10 23 cm -2 ) which hampers the detection of the WAs in the soft X-
rays. For the UFOs, we consider the parameters reported in Papers II
and III.
In the following, we use the definition of the ionization parame-
ter £ = L ion /« r 2 (Tarter, Tucker & Salpeter 1969) in which L ion
is the ionizing luminosity between 1 and 1000 Ryd (1 Ryd =
13.6 eV), ft is the number density of the absorbing material and
r is the distance from the central source. Often, different WA com-
ponents with diverse ionization, velocities and column densities are
detected for each source and there might be time variability, es-
pecially between observations spaced by several years. Moreover,
there might be some intrinsic inhomogeneities in the density and
ionization structure of the absorbing material. However, here we
do not consider any subtle variations in the WA properties because
we focus on deriving the global properties of the outflows and we
refer the reader to the papers reported in the notes of Table 1 for
more information. Therefore, for each source we report the various
ionized absorption components in Table 1 . If more than one paper
reported the analysis of the same source, we averaged the values
of the components with equivalent parameters. This allows us to
minimize the scatter due to different analysis methods employed
by different authors and the effects of time variability as well. The
WA parameters for all the sources, along with their central black
hole masses and average (absorption corrected) ionizing luminosi-
ties L ion derived from the XMM-Newton observations reported in
Paper I, are reported in Table 1. In the subsequent correlation anal-
ysis we will include the points from these multiple zones of WAs
separately.
For comparison, we use the parameters of the highly ionized Fe
K- shell absorbers reported in Paper II. These were initially simply
distinguished as UFOs or non-UFOs if their velocity was higher
or lower than 10 000 km s -1 , respectively. However, as noted in
Papers II and III, their parameter distributions are not bimodal, but
they actually show a roughly continuous distribution in ionization,
column density and velocity, with the UFOs lying at the more
extreme side. This point is also confirmed by the analysis of Gofford
et al. (2012). Bearing this in mind, in the following we will continue
to refer to them as UFOs and non-UFOs (to indicate those detected
in the Fe Xband with an intermediate ionization/velocity), but they
will be correctly considered together in the subsequent correlation
analysis in Section 4.
The number of sources having papers reporting the detection
of WAs is 21/35, therefore the fraction of objects with WAs is at
least >60 per cent, consistent with previous studies (Reynolds 1997 ;
George et al. 1998; Blustin et al. 2005; McKernan et al. 2007; Winter
2010). If we consider the sources showing absorbers in the form
of WAs, UFOs or non-UFOs this increases to 26/35, >74 per cent.
This suggests that the absolute majority of bright Seyfert 1 galaxies
do show some form of ionized X-ray absorption if examined with
sufficiently high signal-to-noise ratio (S/N) observations (Winter
2010; Paper I). The fraction of sources showing UFOs is 12/35
(>34 per cent) and 8/12 (>67 per cent) of these show also WAs. If
we consider the UFOs and non-UFOs together we obtain a fraction
of 16/35 (>46 per cent) and consequently 11/16 (>69 per cent)
of these sources show also WAs. These fractions might possibly
depend also on the inclination of the flow with respect to the line
of sight. Considering the fact that some absorbers may have not
been detected due to low S/N, variability or simply because some
sources have no grating observations or the WAs were not studied
in detail, we emphasize that these fractions do represent only lower
limits.
3 WARM ABSORBER PARAMETERS
We estimate the lower and upper limits of the distance, mass outflow
rate and kinetic power of the WAs following the method outlined in
Paper III. An upper limit on the line of sight projected location can
be derived from the definition of the ionization parameter reported
in Section 2 and the requirement that the thickness of the absorber
does not exceed its distance to the black hole, A H — ftAr < nr (e.g.
Crenshaw & Kraemer 2012):
Lnax = L[ on / N h, (1)
the material cannot be farther away than this given the observed
ionization and column. Instead, an estimate of the minimum dis-
tance can be derived from the radius at which the observed velocity
corresponds to the escape velocity:
Lnin = 2GM B H/^ouf (2)
For the calculation of the mass outflow rate we use the expression
derived by Krongold et al. (2007), which is more appropriate for a
biconical wind-like geometry:
Af out = f(8, (p)7T/xm p N R v 0Ut r, (3)
where f(8, 0) is a function that depends on the angle between the
line of sight to the central source and the accretion disc plane, 8 ,
and the angle formed by the wind with the accretion disc, 0 (see fig.
12 of Krongold et al. 2007). Instead, /x = ft H /ft e — 1/1.4 for solar
abundances. For a roughly vertical disc wind (0 ~ 7t/2) and an
average line-of-sight angle of 8 ~ 30° for the Seyfert Is considered
here (Wu & Han 2001) we ha ve/(<$, 0) ~ 1.5. Full details on the
derivation of this formula can be found in appendix 2 of Krongold
et al. (2007).
This expression for the mass outflow rate has also the important
advantage of not relying on the estimate of the covering and filling
factors. This is due to the fact that it takes into account only the net
observed thickness of the gas, allowing for clumping in the flow.
Thus, there is not the need to include a linear (or volume) filling
factor, since we are interested in estimating the net flow of mass,
starting from the observed column density and velocity. Moreover,
the covering factor is implicitly taken into account by the function
f(8 , 0) when calculating the area filled by the gas, constrained
between the inner and outer conical surfaces. The assumptions are
that the thickness of the wind between the two conical surfaces
is constant with 8 and that this is smaller than the distance to the
source. However, as already noted in Paper III, we obtain equivalent
results including a dumpiness factor of r/r along the line of
sight in the spherical approximation case and using a covering
fraction Cf ~ 0.2 f(8, 0) ~ 0.4, which is consistent with observations
(e.g. Blustin et al. 2005; McKernan et al. 2007). Moreover, this
expression has the same parameter dependencies as that recently
employed by Crenshaw & Kraemer (2012). Equation (3) is actually
more conservative, yielding mass outflow rates that are roughly a
factor of 2 lower.
Neglecting additional acceleration of the outflow, i.e. assuming
that it has reached a constant terminal velocity, the kinetic (or me-
chanical) power can consequently be derived as
Ek = \M OM vl v (4)
The estimates of these parameters are reported in Table 1. We
also calculated the outflow momentum rate as P out = M out u out and
subsequently compared it to the momentum flux of the radiation
field, Prad = Lboi/c. The bolometric luminosity L bo i is estimated as
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
4 F. Tombesi et al.
Table 1 . Parameters of the soft X-ray WAs for the type 1 Seyferts in the sample.
Source
logM B H
logL E dd
(MQ/erg s _1 )
Obs
log^ion
(erg s -1 )
log?
(erg s -1 cm)
log
(cm -2 )
log ^out
(cm s -1 )
logr max
logrmin
(cm)
iogM 0 T
log
(g s -1 )
log^
log£]f n
(erg s -1 )
NGC 4151*
7.1 30 /45.2
C 1
42.9
-2.50
-22.50
-8.00
17.9/17.5
25.2/24.8
40.9/40.5
IC4329A*
8.1 31 /46.2
c 2 ,x 3
44.1
-1.37+0.06
21. 12+0.01
24.4/...
.../...
.../...
0.38 ± 0.07
20.94 ± 0.04
<7.20
22.6/20.1
27.7/25.1
41.8/39.2
2.06 ± 0.05
21.49 ±0.05
<7.30
20.5/19.9
26.1/25.5
40.4/39.8
NGC 3783f
7.5 30 /45.6
c 2
43.5
0.40 + 0.10
21.30 + 0.04
7.74^ ± 0.02
21.8/18.4
27.6/24.3
42.8/39.4
2.10 + 0.10
21.78 ±0.07
7.70 ± 0.01
19.6/18.5
25.9/24.8
41.0/39.9
2.95 ± 0.07
22.00 ± 0.06
7.88 ± 0.01
18.5/18.1
25.3/24.9
40.9/40.4
MCG+8-11-11
7.2 32 /45.3
X 4
43.9
2.66 ± 0.20
22.04 ± 0.24
19.1/...
.../...
.../...
NGC 5548
7.8 30 /45.9
c 2
43.8
2.20 ± 0.20
20.78 ± 0.24
<7.75
20.8/18.7
26.1/24.0
41.3/39.2
NGC 3516f
7.2 35 /45.3
c 2
43.7
2.40 ± 0.20
21.48 + 0.14
7.96 + 0.04
19.8/17.7
26.0/24.0
41.7/39.6
NGC 4593
6.7 30 /44.8
c 2
43.3
2.40 ± 0.20
21.30 ±0.22
<7.00
19.6/17.1
24.7/24.2
38.4/37.9
Mrk 509*
8.1 30 /46.2
c 5
44.3
1.76 + 0.14
21.31 ±0.09
7.44 + 0.10
21.2/19.7
26.8/25.2
41.4/39.8
MCG-6-30-15
6.2 32 /44.3
c 2
43.7
0.20 + 0.10
21.60 + 0.11
22.9/...
.../...
.../...
2.10 + 0.10
21.48 ±0.14
20.1/...
.../...
.../...
3.70 ± 0.20
22.48^;g
8.19 + 0.02
17.5/16.3
25.0/23.7
41.1/39.8
Ark 120*
8.2 30 /46.3
X 6
44.5
Mrk 110
7.4 30 /45.5
X 7
44.3
NGC 7469
7.1 30 /45.2
X 8
43.6
On
0
1 +
o o
o o
20.18 + 0.26
8.00 ± 0.22
21.7/17.5
26.8/22.6
42.5/38.3
IRAS 05078+1626
6.9 33 /45.0
X 9
43.6
2 50 +1 '°°
0.40
24.11 ±0.07
17.0/...
.../...
.../...
Mrk 279f
7.5 30 /45.6
c 10
44.1
0.47 ± 0.07
20.09 ± 0.08
7.31 ±0.11
23.6/19.3
27.7/23.4
42.0/37.7
2.49 ± 0.07
20.51 ±0.11
7.75 ±0.10
21.1/18.4
26.1/23.5
41.3/38.7
NGC 5 26 A
6.2 33 /44.3
43.6
NGC 3227
7.6 30 /45.7
c 2 ,x n
42.1
1.21 ±0.10
21.04 ±0.04
7 62 +0-27
/ - OZ -0.12
19.9/18.8
25.3/24.2
40.3/39.1
2.90 + 0.15
21.38^;22
8.31 + 0.03
17.8/17.4
24.3/23.9
40.7/40.2
NGC 7213
8.0 33 /46.1
X 12
42.6
ESO511-G30
43.7
Mrk 79*
7.7 30 /45.8
X 13
43.9
-1.20
-21.00
21.7/...
.../...
.../...
NGC 4051*
6.3 30 /44.4
c 2 ’ 14 ,x 15
42.2
-0.86
20.49 ± 0.08
7.26 + 0.15
22.6/18.2
27.1/22.8
41.3/37.0
0.60 + 0.16
20.18 + 0.08
7.34 ± 0.06
21.4/18.0
25.7/22.4
40.1/36.8
1.85 + 0.08
20.59 ± 0.09
7.79 ± 0.03
19.8/17.1
24.9/22.3
40.2/37.6
2.78 + 0.17
21.28 + 0.08
7.74 ± 0.02
18.1/17.2
24.0/23.1
39.1/38.2
3.35 ± 0.04
22.33 ± 0.04
8.62 ± 0.01
16.5/15.5
24.3/23.2
41.2/40.2
Mrk 766*
6.1 32 /44.2
c 2
43.3
2.00 + 0.10
2°.30^; 43
21.0/...
.../...
.../...
3.10 + 0.20
20.78^22
19.4/...
.../...
.../...
Mrk 841*
7.8 33 /45.9
X 16
43.9
1.80 + 0.11
21.39 ±0.13
-7.00
20.7/20.2
25.9/25.4
39.6/39.1
3.10 + 0.23
22.27 ± 0.22
-8.00
18.5/18.2
25.6/25.3
41.3/41.0
Mrk 704
7.6 33 /45.7
X 17
43.8
i ^'7+0.27
A,Z/ -0.52
20.30 ±0.11
8.13 + 0.10
22.2/17.8
27.4/23.0
43.4/39.0
2.70 ± 0.30
20.43 ±0.15
7.73 ± 0.08
20.6/18.6
25.6/23.6
40.8/38.8
Fairall 9
8.4 30 /46.5
C 2 ,X 18
44.1
ESO 323— G77f
7.4 33 /45.5
X 19
44.0
1H 0419-577*
8.6 32 /46.7
X 20
44.6
Mrk 335
7.2 30 /45.3
X 21
44.1
ESO 198-G024
8.3 33 /46.4
X 22
44.0
Mrk 290*
7.7 33 /45.8
C 23 ,X 23
43.6
2.43 ± 0.01
21.69 + 0.03
7.65 ± 0.02
19.5/18.8
25.7/25.0
40.7/40.0
Mrk 205*
8.6 34 /46.7
X 24
44.2
Mrk 590
7.7 30 /45.8
C 25 ,X 25
43.3
H 0557-385
7.6 33 /45.7
X 26
43.4
0.50 + 0.18
21.30 ±0.13
21.6/...
.../...
.../...
2.33 ± 0.03
22.11 ±0.03
19.0/...
.../...
.../...
TON SI 80
7.1 32 /45.2
c 27
44.2
PG 1211 + 143*
8.2 30 /46.3
X 28
44.3
Ark 564
6.1 32 /44.2
C 2 ,X 29
44.6
-0.86 + 0.10
19.95 ± 0.09
<7.00
25.5/18.5
29.3/22.3
43.0/36.0
0.87 ± 0.07
20.38 ± 0.04
<7.00
23.3/18.5
27.5/22.7
41.2/36.4
2.58 ± 0.05
20.54 + 0.17
<7.18
21.5/18.2
26.0/22.7
40.1/36.7
Notes. C and X stand for grating observations with Chandra or XMM-Newton, respectively. The * and f mark the sources with detected Fe K absorbers
identified with UFOs and non-UFOs in Paper II, respectively. 1 Kraemer et al. (2005); 2 McKernan et al. (2007); 3 Steenbrugge et al. (2005); 4 Matt et al.
(2006); 5 Yaqoob et al. (2003); 6 Vaughan et al. (2004); 7 Cardaci et al. (2011); 8 Blustin et al. (2003); 9 Svoboda, Guainazzi & Karas (2010); 10 Costantini et al.
(2007); 1 1 Markowitz et al. (2009); 12 Starling et al. (2005); 13 Gallo et al. (2011); 14 Lobban et al. (2011); 15 Pounds & Vaughan (2011); 16 Longinotti et al.
(2010); 17 Laha, Dewangan & Kembhavi (2011); 18 Emmanoulopoulos et al. (2011) 19 Jimenez-Bailon et al. (2008); 20 Pounds et al. (2004); 21 Gondoin et al.
(2002); 22 Porquet et al. (2004); 23 Zhang et al. (2011); 24 Reeves et al. (2001); 25 Longinotti et al. (2007); 26 Ashton et al. (2006); 27 Rozariska et al. (2004);
28 Pounds et al. (2003); 29 Smith, Page & Branduardi-Raymont (2008); 30 Peterson et al. (2004); 31 Markowitz (2009); 32 Bian & Zhao (2003a); 33 Wang &
Zhang (2007); 34 Wandel & Mushotzky (1986); 35 Onken et al. (2003).
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies 5
Table 2. Results of the linear regression and partial correlation analysis for the measured absorbers parameters.
X
y
a
Dev(a)
b
D Q\(b)
Scatt
Rp
dof
P null
Z
a k
Ppart
All absorbers together
log£
logN H
0.72
0.12
20.00
0.33
0.71
0.73
55
1.0 x 10 -10
log t+ut
0.5
0.09
1.0 X 10" 10
log£
log ^out
0.65
0.10
6.61
0.35
0.52
0.79
45
4.0 x 10“ u
logN H
0.3
0.07
1.0 X 10“ 4
logN H
log l+ut
0.69
0.09
-6.84
1.97
0.45
0.77
65
3.0 x 10" 14
log£
0.3
0.09
5.0 x 10" 4
WAs
log£
logN H
0.73
0.13
19.92
0.29
0.83
0.50
1.5 x 10" 3
log£
log r’out
0.31
0.08
7.19
0.18
0.30
0.69
1.0 x 10“ 3
logN H
log l+ut
0.48
0.21
-2.46
4.47
0.30
0.52
2.0 x 10“ 2
UFOs and non-UFOs
log£
logN H
0.62
0.53
20.62
1.98
0.27
0.41
9.0 x 10“ 2
log£
log l+ut
0.63
1.09
6.72
4.46
0.60
0.23
2.6 x 10" 1
logN H
log l+ut
1.43
1.30
-23.74
30.10
0.79
0.29
2.9 x 10 _1
Notes, a and b are the slope and intercept of the linear correlation fits with standard deviations dev(a) and dev(Z?), respectively. Scatt represents
the internal scatter in units of dex. Rp is the Pearson coefficient, dof is the number of degrees of freedom. P nu ii is the probability of the null
hypothesis that there is no correlation between x and y. z is the third variable against which the partial correlation analysis is performed, tk and
<tk are Kendall’s partial correlation coefficient and the variance, respectively. P part is the null hypothesis probability of the partial correlation.
L bo i = kbo\L\on, where /+oi — 10 (McKernan et al. 2007; Vasudevan
& Fabian 2009; Lusso et al. 2010).
4 CORRELATION ANALYSIS
4.1 Measured absorber parameters
Initially, we consider the data from all WAs, UFOs and non-UFOs
together. We compare their ionization parameter, velocity and col-
umn density and search for possible correlations. The large dy-
namic range of ~6 orders of magnitude in ionization, ~4 in column
density and ~3 in outflow velocity allows, for the first time, to in-
vestigate the general relations and trends among these parameters.
We fit the power-law model logy = a logx + b to the data sets
using the bivariate correlated errors and intrinsic scatter (BCES)
regression method (Akritas & Bershady 1996) which takes into
account measurement errors in both the ‘X’ and T coordinates
as well as the intrinsic scatter. This method has been widely used
in fitting data sets in the astronomical community (e.g. Sani et al.
2011). No a priori dependent variable in the fitting is assumed and
we treat the variables symmetrically. Uncertainties on the param-
eters derived from the fits are estimated after carrying out 10 000
bootstrap resamples of the data. In order to estimate the signifi-
cance of these correlations we calculated the Pearson coefficient
Rp.
The results of the fits are listed in Table 2 and the scatter plots and
best- fitting relations are shown in Fig. I. 1 All three correlations have
a high statistical significance (>6<r). We note that the wide dynamic
range covered by the observed absorption components allows us to
fill the whole parameter space, with the WAs and the UFOs at the
two sides.
In particular, the slope 0.72 =b 0.12 of the correlation between
the ionization parameter and column density in Fig. 1(a) suggests
1 In calculating the correlations we used only the points constrained within
their errors of measure. Instead, in the figures we plot also those with upper
or lower limits for completeness.
that the column density is higher for more ionized absorbers. 2 From
Fig. 1(b) we observe a linear relation with slope 0.65 =b 0.10 be-
tween the ionization and velocity, indicating that the faster outflows
are also more ionized. Finally, Fig. 1(c) shows a relation between
column density and velocity with slope 0.69 =b 0.09, which indi-
cates that the faster outflows have also higher columns. In general,
these correlations suggest that the faster the outflow, the higher the
column density and ionization parameter. The possible systematics
and selection effects affecting these relations are discussed later in
Section 4.3.
In order to test whether the correlations between two parameters
are driven by their mutual dependence on a third one, we then
performed also a partial correlation analysis (Akritas & Siebert
1996). We quantify this effect using Kendall’s partial correlation
coefficient, t k , which takes into account also censored data. 3 The
results of the partial correlation analysis are reported in the last three
columns of Table 2. This test indicates that the correlations between
the different parameters are only marginally interdependent.
We performed also a correlation analysis considering separately
the WAs and the other absorbers (UFOs and non-UFOs). As already
noted in Section 2, the distinction between UFOs and non-UFOs
is only arbitrary and based on their velocity being higher or lower
than 10 000 km s _1 . In fact, their parameter distributions reported
in Papers II and III are not bimodal. This test was done in order
to check if there are significant differences when considering the
two populations separately. The results are reported in Table 2. Due
to the large error bars, smaller number of data points and reduced
dynamic range of the two separated populations, their parameter
values are less constrained and the significance is lower, especially
for the highly ionized absorbers. However, a possible difference
can be noted in the best-fitting value of the slope of the relations
between log£ — log u 0Ut and log/V H — log u 0Ut , one of the highly
2 In Fig. 1(a) we note a possible WA outlier with high column density,
Ah ~ 10 24 cm -2 , and a relatively high ionization, log£ ~ 2.5 erg s -1 cm.
As reported in Table 1, this corresponds to the Seyfert 1.5 galaxy IRAS
05078+1626. However, given that it has not a velocity estimate, this is not
considered anymore in all the successive relations.
3 The partial correlation analysis takes into account only the points con-
strained within their errors of measure and those with upper limits.
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
6 F. Tombesi et al.
log N„ (cm 2 )
Figure 1 . Correlations for measured outflow parameters. Scatter plots of
log £ versus logNn (panel a), log £ versus log i> ou t (panel b) and logNn versus
log ^out (panel c) for the WAs (red), non-UFOs (green) and UFOs (blue).
The arrows indicate the lower or upper limits. The solid lines represent the
best-fitting linear correlations and the grey shadowed areas indicate the 2 o
confidence bands.
ionized absorbers being steeper, although they are still consistent
given the large uncertainties. This point will be addressed later in the
discussion section. Given the limited quality of the current data sets
we are not in a position to strongly constrain possible differences
between the separated correlations.
Finally, we note that Blustin et al. (2005) attempted a similar
correlation analysis of the WAs collecting their parameters from
the literature. They considered a sample of 23 Seyferts and quasars,
finding useful information for only 14 of them. They found a sig-
nificant correlation between log£ and logA/ H , as reported also by
other authors (Holczer, Behar & Kaspi 2007; Behar 2009). Exclud-
ing instead the few fast and highly ionized outflows reported at that
time, they found only a very marginal correlation between log % and
log ^out and no significant correlation between log w 0 ut and log JV H -
However, it is important to bear in mind that the inability to find
a significant correlation does not necessarily exclude its existence.
In fact, after more than seven years of additional observations and
analysis, we can now define a larger sample of sources (35) with
higher S/N observations, which allowed us to increase the number
and quality of the reported WA components and, thus, better con-
strain the correlations between the parameters. The inclusion of the
UFOs also allows us to significantly expand the parameter space.
4.2 Derived outflow parameters
We checked for possible correlations/trends among the derived out-
flow parameters of the WAs listed in Table 1 and those of the UFOs
and non-UFOs reported in Paper III. Given that we can only estimate
upper and lower limits, in order to carry out the fits we performed
10 000 Monte Carlo simulations for each data set, considering a
random value between the lower and upper limits for each point
(assuming a uniform distribution in log space). We then calculated
the linear regression and the corresponding statistics.
The average black hole mass of the sources in the sample is
log(M BH /MQ) ~ 7.4 with a standard deviation of ~0.7 dex. The
difference is not large, but in order to take into account the expected
scaling of the outflow parameters with mass, we normalized the
distance to the Schwarzschild radius r s = 2GM BH /c 2 , the kinetic
power to the Eddington luminosity L Edd = An GM E}d m p c / a T —
1.26 x 10 38 (M bh /Mq) erg s -1 , the momentum to the Eddington
momentum rate P Edd = L Edd /c and the mass outflow rate to the
Eddington rate M Edd = r]L Edd /c 2 , where r) ~ 0.1.
The parameters and significance of the correlations are listed in
Table 3. The plots are reported in Figs 2-4. 4 These show that, even
if the difference between the lower and upper limits can be large in
some cases, the wide dynamic range of the parameters still allows
us to estimate significant correlations/trends among them. The plots
of the velocity with respect to the kinetic power and mass outflow
rate are reported in Fig. 2. In Fig. 2(a) we note a strong trend of
increasing the kinetic power for increasing velocity, with a positive
slope of 1.77 ± 0.14. This is close to the trend expected for a
roughly constant mass outflow rate. In fact, from Fig. 2(b) we note
that, besides the large uncertainties, M out does not vary much with
v out , although there is a possible weak trend of decreasing mass
outflow rate with increasing velocity with slope —0.21 =L 0.14 (see
Table 3).
4 For each of these relations we considered only the points constrained
between the errors or upper/lower limits in both the X and Y axes. Therefore,
the number of points can slightly differ from one plot to the other.
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies
7
Table 3. Linear regression results for the derived outflow parameters.
x y a dev(a) b dev(&) Rp /Lull
^out
log(F K / T E dd)
log v out
log(M 0 ut/A/Edd)
log (r/r s )
log£
log (r/r s )
log A h
log (r/r s )
log u ou t
log (r/r s )
log(A/ ou t/ A^Edd)
log (r/r s )
log(P 0 ut/ /*Edd)
log (r/r s )
log(F K / F E dd)
log / rad
log Pout
1.77
0.14
- 18.89
- 0.21
0.14
1.24
- 0.58
0.04
5.80
- 0.44
0.04
24.21
- 0.40
0.03
10.47
0.16
0.07
- 1.29
- 0.22
0.07
- 0.44
- 0.60
0.09
- 0.89
0.76
0.19
8.68
1.19
0.89
3.8 x 10“ 15
1.19
- 0.22
1.6 x lCT 1
0.21
- 0.85
8.0 x 10“ 16
0.20
- 0.86
1.3 x 10" 13
0.11
- 0.89
1.6 x 10" 15
0.32
0.38
5.0 x 10“ 3
0.33
- 0.45
7.2 x 10“ 4
0.38
- 0.73
3.1 x 10“ 10
6.41
0.56
1.0 x 10“ 2
Notes, a and b are the slope and intercept of the linear correlation fits with standard deviations
dev(a) and dev(fr), respectively. Rp is the Pearson coefficient. /Lull is the probability of the null
hypothesis.
Figure 2. Comparison between the outflow velocity log i> ou t> the outflow kinetic power log(ZsK/TEdd) (panel a) and the mass outflow rate log(M out /AfEdd)
(panel b), respectively. The points relative to the WAs (red open circles), non-UFOs (green filled triangles) and UFOs (blue filled circles) are reported. The
error bars indicate the upper and lower limits and the points are the average between the two. The solid lines indicate the best-fitting linear regression curves.
Instead, in Fig. 3 we show the plots of the different parame-
ters with respect to the line of sight projected distance in units of
r s . From panels a, b and c we note very significant trends of de-
creasing the ionization parameter, column density and velocity with
distance, with slopes of —0.58 =b 0.04, —0.44 ± 0.04 and —0.40 =b
0.03, respectively (see Table 3). In particular, we checked that the
correlation between the velocity and distance is not an induced re-
lation from the fact that the lower limits have been estimated using
equation (2) in Section 3, i.e. assuming that the observed velocity
is equivalent to the escape velocity, as this relation is independently
confirmed using the upper limits alone derived from equation (1), as
discussed also later in Section 4.3. In panels d and e we can observe
a weak increase/decrease of the mass outflow rate/momentum rate
for increasing distance, with slopes of 0.16 =b 0.07 and —0.22 =L
0.07, respectively. Instead, in panel f we can note a more pronounced
and significant trend of increasing the observed outflow mechanical
power with decreasing distance, going from the WAs to the UFOs,
with a slope of —0.60 =b 0.09.
It should be noted that the whole parameter space is essentially
uniformly filled, with distances ranging from ~100 r s from the
black hole up to ~kpc scales. The UFOs occupy the lower end of the
distribution at the smaller distances, where the ionization, column
and velocity are higher. In particular, extrapolating the relations
reported in Table 3 and Fig. 3 to the innermost possible radii of the
order of log(r/ r s ) ~lwe obtain that the ionization of the gas reaches
very high values of log£ ~ 5-6 erg s -1 cm, the column becomes
mildly Compton-thick, Ah — 10 24 cm -2 , and the outflow velocity
approaches significant fractions of the speed of light. The values of
the parameters gradually decrease with increasing distance, going
from the UFOs to the WAs.
Finally, Fig. 4 shows a significant (99 per cent confidence level)
linear relation between the radiation momentum rate, P rac i = L bo \/c,
and the outflow momentum rate of the UFOs, P out , with a slope of
0.76 =b 0.19. The possible physical implications of this and the
previous relations will be briefly addressed in discussion Section 5.
4.3 Possible systematics and selection effects
In the calculation of the correlations in the previous sections we
took into account the uncertainties in the ionization parameter, col-
umn density and outflow velocity as reported in Table 2. However,
when considering such a large data set and especially when collect-
ing results from the literature, it is important to bear in mind the
existence of possible systematics and selection effects.
As already discussed in Paper II, different assumptions of the
velocity broadening of the lines can generate some variations in
the estimated column density. Regarding the UFOs, in Paper II
we already took into account this effect testing different velocity
broadenings for the lines that were not resolved and included the
larger error bars. For the column densities of the WAs collected
from the literature, this effect is marginal given that the estimates
were derived using high energy resolution data. Another parameter
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
8 F. Tombesi et al.
w
O)
a5,
uj'
o
§
■ CL
o
log(r/g
Figure 3. Comparison between the line-of-sight projected distance log (r/r s ) (in units of the Schwarzschild radius) and the ionization parameter log £ (panel
a), the column density log Ah (panel b), the outflow velocity log v out (panel c), the mass outflow rate log(M 0 ut/^Edd) (panel d), the outflow momentum rate
log (/^ut/ f^Edd) (panel e) and the outflow kinetic power \og(E^/ L^dd) (panel f), respectively. The points relative to the WAs (red open circles), non-UFOs
(green filled triangles) and UFOs (blue filled circles) are reported. The solid lines indicate the best-fitting linear regression curves. Assuming the typical black
hole mass of Mbh — 10 7 Mq, the distance scale can be easily converted from the Schwarzschild radii to parsec considering that lpc ~ 10 6 r s .
that can affect the estimate of the column density is the assumed
elemental abundance. However, as already discussed in Paper III,
even allowing for a factor of 2 difference with respect to the stan-
dard solar values, the discrepancy in the column density is <0.2
dex, within the typical errors of measure. The fact that the column
densities do not exceed the value of A/ H — 10 24 cm -2 , especially
noticeable in Fig. 1(c), could in principle be affected by the fact
that the photo-ionization code Xstar cannot treat Compton-thick
absorbers. However, the data do not seem to require significantly
higher columns as a good spectral modelling of the highly ionized
UFOs in Paper II was already obtained with columns in the range
N n = 10 22 -10 24 cm- 2 .
It is known that the ionization parameter log £ has a dependence
on the assumed ionizing continuum. In Paper III we estimated that
the possible uncertainty on the continuum slope may induce a max-
imum systematic error of ~0.4 dex on the ionization parameter,
within the typical internal scatter of the relations shown in Fig. 1
(panels a and b). Moreover, the large range in observed ionization
states reduces the importance of this effect when performing the
linear regression fits.
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies 9
Figure 4. Outflow momentum rate of the UFOs with respect to the mo-
mentum rate of the radiation field. The error bars indicate the upper and
lower limits and the points are the average between the two. The solid line
indicates the best-fitting linear regression curve.
Considering the column density and ionization in Fig. 1(a), there
might be some possible selection effects at the two ends of the
distribution. In fact, the limited S/N of the observations could have
hampered the detection of the weak spectral features from absorbers
with high ionization and low column density. On the other hand, the
fact that we can observe WAs only for type 1 sources could have
limited the relevance of lowly ionized/neutral absorbers with high
column densities. However, these latter types of absorbers, usually
found only in type 2 sources, have intrinsic velocity consistent with
zero and probably have a different origin than the AGN outflows
studied here, such as the ~pc scale molecular torus or large-scale
dust lanes in the host galaxy itself. The combination of these two
possible selection effects could have induced a slight steepening of
the log$-logN H relation and of the radial density profile succes-
sively discussed in Section 5.2. Moreover, these estimates do not
take into account the possible presence of additional fully ionized
material, which does not imprint any observable spectral absorption
features.
We note that another important parameter of these outflows is
their inclination with respect to the line of sight. Unfortunately, the
estimate of this parameter is not well constrained for each source
yet, but the typical inclination of type 1 Seyferts is ~30°, with a
range between ~10° and 60° (Wu & Han 2001). Therefore, this
ensures that the difference among the sources of our sample is not
large. However, this might contribute to some of the internal scatter
observed in Fig. 1. Anyway, given the large dynamic range, small
differences between sources do not significantly affect the derived
scale relations.
As already discussed in Paper III, the expression for the mass out-
flow rate used in Section 3 introduces a possible systematic source
of uncertainty from the assumed typical inclination and opening
angle of the wind, which however is constrained within a factor
of ~0.3 dex (Krongold et al. 2007). Instead, the typical systematic
uncertainty on the black hole masses derived using the reverberation
mapping technique is <0.5 dex (Peterson et al. 2004).
Regarding the outflow velocity of the absorbers, as already dis-
cussed in Paper II, we might be losing some of the components with
the highest blue shift due to the fact that they are usually detected in
the Fe K band at E > 7 keV where both the energy resolution and
effective area of the EPIC-pn instrument on board XMM-Newton
are worse and the detector has essentially no sensitivity above
E~ 10 keV.
We also checked that the methods used to estimate the lower and
upper limits of the distance of the absorbers reported in equations
(1) and (2) do not introduce significant systematics in the relations
reported in Figs 2 and 3. For instance, the same dependence between
the velocity and distance with a slope consistent with 0.4 shown
in Fig. 3(c) is independently found also when considering only the
upper limits of the distance derived from the ionization parameter
in equation (1).
All these sources of uncertainties can contribute to the signifi-
cant internal scatter of >0.5 dex observable in the plots in Fig. 1.
Moreover, we should bear in mind also the possibility that inhomo-
geneities and variability of the absorbers and also the distinct anal-
ysis methods employed by the different authors could contribute
to the observed scatter as well. A direct, more homogeneous and
systematic spectral analysis of the WAs, similar to what we have
done for the UFOs in Papers I and II, would reduce the importance
of these possible systematics, but this is beyond the scope of the
present paper. Here we focus more on the global picture, noting also
that very detailed spectral analysis of WAs could be performed only
on a handful of sources as they require very large and high-quality
data sets.
Finally, here we tested for simple linear relations in log-space
between the absorber parameters, such as ionization, column den-
sity, outflow velocity, distance and energetics. This represents a
good first order approximation but there might be slight variations
in the slope, particularly at the low and high ends of the distribu-
tions, which could indicate different regimes. This possibility will
be briefly considered later in the Discussion section.
5 DISCUSSION
5.1 Unification of the X-ray outflows
In all the tests considered and in Figs 1-3 we find that the WAs,
non-UFOs and UFOs, within the observational uncertainties, show
similar relations between their parameters. In particular, the WAs
and UFOs lie always at the two ends of the distributions, with the
intermediate non-UFOs in between, and they roughly uniformly
cover the whole parameter space. When considered all together,
we find very significant correlations between their parameters, such
as ionization, column density, velocity and distance. These results
strongly suggest that these absorbers, sometimes considered of dif-
ferent type, could actually represent parts of a single general strati-
fied outflow observed at different locations along the line of sight.
If they were unrelated, the points relative to the different absorbers
would be mixed and would not display any significant correla-
tions, especially when considering several sources together. A sim-
ple schematic diagram of a stratified accretion disc wind is shown
in Fig. 5.
The stratification along the line of sight can easily explain the
relations in Figs 1-3. For instance, the fact that the faster absorbers
are also more ionized suggests that we see components that are
ejected closer to the black hole. The increase in column density for
higher ionization can be explained with a negative gradient with
distance, as also discussed in the next section when estimating the
density profile.
The observed correlation between the velocity and the distance
shown in Fig. 3(c) and reported in Table 3 is consistent with an
approximate slope of —0.40 =b 0.03. This is essentially equivalent to
the one expected for a wind that has an outflow velocity proportional
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
10 F. Tombesi et al.
I I I I I
Figure 5. Simple schematic diagram of a stratified accretion disc wind. The
figure is not to scale. This is similar to fig. 8 of Kazanas et al. (2012). The
torus, not directly outlined here, may represent an extension of the outer
accretion disc itself.
to the escape velocity at a given location and, therefore, that is able
to escape the system and not fall back. Thus, v out =f w v e sc > where
u esc = (2 GM/r) l/2 is the escape velocity and / v > 1. Therefore, such
type of winds are expected to roughly follow a relation u 0Ut a r _1/2 .
This relation is satisfied in two circumstances: (i) the wind ob-
served at each radius was launched relatively close to that radius
with a velocity proportional the local escape speed or (ii) the wind
was launched at small radii with a velocity close to the escape speed
and then decelerated under the sole effect of gravity, i.e. ballisti-
cally. This latter case is probably unphysical because the wind is
not expanding in pure vacuum, but there is always some interstellar
medium.
In this regard, the observed relation excludes the scenario of a
wind being launched from small radii with a constant terminal ve-
locity, which would lead to a flat profile with radius. However, this
last hypothesis is, at least partially, ruled out unless some decelerat-
ing processes are present, such as shocks or entrainment of some of
the surrounding material. Actually, the fact that the observed profile
is 0.4 instead of exactly —0.5 suggests that some form of inter-
action with the external medium might be present at ^>pc scales.
This is supported by the slight increase of the mass outflow rate in
Fig. 3(d) at large distances of >100 pc, which might indicate some
entrainment of surrounding material. This point will be addressed
also in Section 5.4.
As already discussed in Paper III, the location and charac-
teristics of the UFOs are indeed consistent with accretion disc
winds/outflows and the possible direct connection with the other
non-UFOs and WAs reported here indicates that actually all of
them could be identified with the same global outflow observed at
different locations along the line of sight, the WAs being ejected
at larger distances, of the order of the outer disc and/or torus, the
latter possibly representing a natural large-scale extension of the
disc itself (see Fig. 5). In fact, the correlation analysis and the plots
in Figs 1-3 point towards an underlying connection among these
ionized absorbers.
It will be discussed later in Section 5.3.2 that case (i), in which
the wind is launched on a wide range of radii with a velocity pro-
file proportional to the escape velocity, is probably the preferred
interpretation and it is directly predicted by stratified MHD ac-
cretion disc wind models. This interpretation is valid at least for
distances 100 pc scales, as additional effects of interaction with
the surrounding medium might be important beyond that (see Sec-
tion 5.4).
Regarding the WAs, we note that the fact that the line of sight
projected location of some of them is of the order of the putative
obscuring torus at the base of the type 1/type 2 unification models
does not necessarily mean that they are produced there, for instance
as inferred by Blustin et al. (2005). The putative torus is also a
large, equatorial structure which is difficult to relate with winds that
are observed at relatively small inclinations along the line of sight
(~30°), as for the Seyfert 1 galaxies discussed here (Wu & Han
2001). We also note that some authors even suggested that the torus
itself could actually be identified with the slower and outer regions
of a global stratified wind. At these locations, dust would be present
at the disc surface and it could be uplifted and become embedded
in the outflow, explaining its presence in some WA observations
(Konigl & Kartje 1994; Elvis 2000; Kazanas et al. 2012).
Finally, we note that the presence of a photo-ionized outflow
extending from the inner regions around the black hole up to ~kpc
scales could be directly related with the ionization cones observed
through emission lines and images in several Seyfert galaxies, some
of them being also part of this sample (e.g. Storchi-Bergmann et al.
2009; Crenshaw & Kraemer 2000; Dadina et al. 2010; Wang et al.
2011a; Paggi et al. 2012).
5.2 Outflow density profile
An important quantity describing these ionized outflows is their
radial density profile. Considering all these absorbers as different
representations of the same phenomenon, we can derive an estimate
of their average global density profile of the form n(r ) a r ~ a .
From the relation between the column density and distance shown
in Fig. 3(b) and reported in Table 3 we have A H = 10^(r/r s ) a cm -2 ,
where a = —0.44 =b 0.04 and b = 24.21 =b 0.20. The column density
can be expressed also as A H = n(r)Ar ~ n(r)r s (r/r s ), with A r ~
r. Combining these two equations we obtain a rough expression
for the density profile as n(r) ~ no(r/r & ) a ~ \ with n 0 = I0 b /r s ~
( 10 2421 /3 x 10 5 )(M B h/Mq) -1 ~5x lO^Mg -1 cm -3 represent-
ing the density normalization and M 8 = M B h/10 8 Mq. Therefore,
substituting the observed value of a ~ —0.4, we derive that the den-
sity profile has a slope a = 1 — a ~ 1.4. Consequently, for a typical
black hole mass of ~10 7 Mq and for an inner radius of log (r/r s ) ~
1, we obtain a density at the base of the outflow of ~10 10 cm -3 .
For comparison, from the relation between the column density
and the ionization parameter shown in Fig. 1(a) and following Hol-
czer et al. (2007) and Behar (2009), an estimate of the radial density
profile can be determined using also the absorption measure distri-
bution, defined as AMD = <7N H /d(log§). The AMD is the absorp-
tion analogue of the emission measure distribution (EMD), widely
used in the analysis of the emission line spectra, and it represents
the distribution of hydrogen column density along the line of sight
as a function of ionization. Given the relation logN H = a\og% +
b in Table 2 with a = 0.72 =b 0.12, this can then be rewritten as
AMD = (10^alnl0)£ a oc § a . Then, the slope of the radial density
profile can be estimated as a = (1 + 2a)/(l + a) with uncertainty
Aa = Aa/(l + a) 2 (Behar 2009). Substituting the observed quan-
tities in Table 2 we obtain a = 1.42 =L 0.04, consolidating further
this important result.
This value is slightly higher than those reported in a more detailed
analysis of the WAs in a sample of five Seyfert galaxies by Behar
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies 1 1
(2009), who nevertheless suggested that a slight increase could
be present for high ionizations. However, it should also be noted
that the possible selection effects for the absorbers with very low
and high ionization and column densities previously discussed in
Section 4.3 could induce a slight steepening of the density profile
estimated in this way. Moreover, these estimates do not take into
account the possible presence of additional fully ionized material,
which does not imprint any observable spectral absorption features.
Therefore, in both cases the density profile is n oc r~ l 4 . As already
noted by Behar (2009), this scaling rules out two simple scenarios,
a constant density flow ( n oc r°) and, on less stringent grounds, a
steady, mass conserving spherical symmetric radial flow similar to
a stellar wind, in which the mass outflow rate, the opening angle
and the wind velocity are all constant ( n oc r~ 2 ). This density profile
suggests that the outflowing gas is more consistent with a conical/
paraboloidal shaped wind instead of a simple spherical shell.
5.3 Acceleration mechanisms
Once we have established the fact that the ionized absorbers can be
unified as part of a single, large-scale outflow, then a fundamental
question follows: what is (are) the main acceleration mechanism(s)?
The limited dynamic range in luminosity for the sources considered
here hampers a detailed study of possible correlations between the
bolometric luminosity and the parameters of the outflows. In fact,
the average bolometric luminosity is logL bo i — 44.5 erg s -1 with a
dispersion of only ~0.8 dex. Moreover, the existence of different
absorber components for each source might render the search for
such correlations non-trivial. However, we can perform some tests.
In order to get some insights into the possible acceleration mech-
anism^) of the outflow, the velocity alone is not a good parameter
to compare with the source luminosity, while the momentum rate
and the kinetic power are better because they also take into account
the mass outflow rate, i.e. the energetics involved.
Considering the Eddington ratios of the Seyferts in our sample
showing X-ray outflows we obtain an average value of ~0.15 with
a dispersion of ~0.6 dex. This interval is too narrow to investigate
for possible correlations/trends between the Eddington ratio and the
other parameters. However, we note that for a given ratio, the UFOs
are always more powerful than the other outflows.
More insights might be derived comparing the outflow and radia-
tion momentum rates. The outflow momentum rate can be expressed
as
Pout = M out v ~ AnCim v nr 2 v 2 , (5)
where Cf is the covering fraction and can assume values between
0 and 1. Instead, the momentum rate of the radiation field was
defined in Section 3 as P v ad = L bol /c. As already introduced in
Section 4.2 and reported in Fig. 4 and Table 3, there is a significant
roughly linear (slope a = 0.76 =b 0.19) correlation between the
radiation momentum rate and the outflow momentum rate of the
UFOs, therefore P out ~ P ra d- In particular, the average value of
the ratio between these two parameters is consistent with unity,
{Pout/ Trad) = 1.6 =b 1 . 1 , indicating a direct proportionality between
these two quantities.
We checked for the existence of a similar relation also for the
WAs but we could not constrain it given the large errors and scatter.
This is due to the fact that their parameters are less homogeneous
and they cover a wider interval of distances compared to the UFOs.
Nevertheless, we can estimate a much lower average ratio between
their outflow and radiation momentum rates of ~0.05, with a stan-
dard deviation of ~0.26 dex.
Therefore, the linear correlation observable in Fig. 4 suggests
that there should be a significant exchange of momentum between
the radiation field and the outflows or, from the relation M acc =
Lboi/^c 2 , that the power of the outflows is at least related to the
accretion rate.
5.3.1 Radiation pressure
As we can see from panels a and b of Fig. 3, even if the inner
part of the flow represented by the UFOs can reach significant
column densities in the range Nu — 10 22 -10 24 cm -2 , this material
is also highly ionized, with possibly the majority of the elements
lighter than iron being fully stripped of their electrons. Therefore,
it is hard, if not impossible, to invoke radiation pressure from UV
absorption lines as the main acceleration mechanism for these flows.
This process might instead be more important for winds in bright
quasars, given their different SED shape and the higher UV emission
compared to Seyferts (e.g. Proga & Kallman 2004).
The alternative scenario in which radiation pressure could provide
significant acceleration to this material even if it is highly ionized is
then through Compton scattering (e.g. King & Pounds 2003; King
20 10a, b). In particular, in this case a direct proportionality between
the momentum rate of the radiation field and that of the outflow
would be expected:
Pout,R — Cf T e P ra d? (6)
where r e is the electron optical depth to Compton scattering and
Cf is the covering fraction of the wind. If this is the dominating
acceleration mechanism, from the linear relation between the UFO
momentum flux and that of the radiation field reported in Fig. 4 we
derive that the product Cft e should be of the order of unity.
From the fraction of sources with detected UFOs in Paper I we
derive that the global covering fraction of these absorbers is in the
range ~0.4-0.6 and therefore we can assume a typical value of Cf ~
0.5. Considering the observed average column density of the UFOs
of Nu — 10 23 cm -2 we can estimate an optical depth r e — ct^Nu —
0.05. The product of these two values is much lower than ~unity
expected from the relation P out ~ P rad . However, it should be noted
that the column density of the UFOs might have been larger dur-
ing the acceleration phase. Extrapolating the relation between the
column density and the distance shown in Fig. 3(b) and reported in
Table 3 we obtain that the column density at the innermost radii of
log (r/r s ) ~ 1 is V H — 10 24 cm -2 . However, this does not take into
account the possible presence of some additional fully ionized ma-
terial that is not visible through X-ray spectroscopy. For instance,
extrapolating the relation in Fig. 3(a) down to log (r/r s ) ~ 1 would
give rise to log£ > 5 erg s -1 cm. Therefore, we obtain r e ~ 1, in-
dicating that the material can be mildly Compton-thick at the base
of the flow. The product C f r e is now more close to unity.
Combining equations (5) and (6) and substituting the previously
reported definitions for the ionization parameter (Section 2), bolo-
metric luminosity and radiation momentum rate (end of Section
3), we can derive a rough dependence of the outflow velocity with
respect to the ionization parameter:
fout,R
kbo\
47 rm p c
1/2
r e 1/2 | 1/2 ;
(7)
therefore, for the radiation pressure case we would expect v out a
t g 1/2 § 1/2 , which is comparable to that found in Fig. 1(b).
Even if the UFOs have a momentum rate comparable to that
of the radiation field, it is important to check if the luminosity of
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
12 F. Tombesi et al.
these sources is actually enough to accelerate the material to the
escape velocity, which is required for such a wind to leave the
system. Therefore, combining equation (7) with the definition of
the Eddington luminosity in Section 4.2 and imposing that the wind
velocity should be equal or higher than the escape velocity at a
particular location (see Section 5.1), we obtain
where k = L ho \/L Edd is the Eddington ratio, A H = nAr ~ nr is the
observed column density and A/ acc is the column density of the gas
during the acceleration phase. From the previously discussed possi-
bility that the column density at the base of the flow [log (r/r s ) ~ 1],
where the majority of the acceleration should take place, is A/ acc ~
10 24 cm -2 and considering the average observed column density of
the UFOs of Nu — 10 23 cm -2 , we estimate that the radiation is ca-
pable to accelerate the wind to the escape velocity if the Eddington
ratio is >0.2. Given that the average Eddington ratio of the sources
considered here is only k ~ 0. 15, we suggest that radiation pressure
might be relevant to accelerate the observed outflows.
The effect of radiation pressure could be increased from multiple
electron scatterings of the continuum photons if the wind opacity
would be higher than one or if the luminosity of these sources would
be closer to Eddington. However, the maximum opacity at the very
base of the wind is only r e ~ 1 and these sources are sub-Eddington.
Regarding this last point, we note that King (2010b) discussed the
possibility that bright AGNs, such as those discussed here, could
actually be closer to Eddington due to uncertainties on the black
hole mass and bolometric luminosity estimates.
Therefore, these evidence indicate that the UFOs may be accel-
erated, at least partially, by the exchange of momentum with the
radiation field through Compton scattering. This is overall consis-
tent with momentum-driven outflow models (King & Pounds 2003;
King 20 10a, b) which actually predict the existence of highly ionized
outflows with velocity ~0. lc and a linear relation between the wind
and radiation momentum rates. One requirement by these models
is that the optical depth should be of the order of unity, which we
find to be possible at the very base of the flow, and that the covering
fraction should be Cf ~ 1. From observations we derive that Cf ~
0.5 (Paper I), which indicates that the outflow has a significant cov-
ering fraction and it is uncollimated, nevertheless it is not spherical,
but probably more consistent with a conical/paraboloidal bipolar
wind-like shape.
5.3.2 Magnetohydrodynamic mechanisms
An additional mechanism that could provide a concurrent accel-
eration for this highly ionized material is represented by MHD
processes (Blandford & Payne 1982; Everett & Ballantyne 2004;
Fukumura et al. 2010; Kazanas et al. 2012). We note that MHD
mechanisms are a fundamental requirement for generating the vis-
cosity in accretion discs (e.g. Abramowicz & Fragile 2011) and they
are postulated as one of the main heating mechanisms of the X-ray
emitting corona (e.g. Haardt & Maraschi 1991). Moreover, several
Seyfert galaxies do show evidence for weak radio jets (Ulvestad &
Wilson 1984; Ulvestad, Antonucci & Barvainis 2005; Wang et al.
201 lb). Therefore, significant magnetic fields might well be present
in the inner regions of these radio-quiet sources.
For instance, Fukumura et al. (2010) studied the ionization struc-
ture of self- similar MHD winds off Keplerian accretion discs in
AGNs. The magnetic field is dragged by the rotating disc plasma,
and as the wind leaves the disc, magnetic torques act on the gas and
the wind is magnetocentrifugally accelerated. In this case, the wind
is found to end up with a terminal velocity roughly a few times
the initial rotational speed. Therefore, a typical feature of these
winds is that they retain a clear information about their launch-
ing region. At large distances (>0.1-1 pc) the putative AGN torus
may actually represent an extension of the outer accretion disc
itself.
Then, Kazanas et al. (2012) generalized this concept and sug-
gested that MHD winds could actually represent the fundamental
structure at the base of the unified model of AGNs. In particular,
they derived simple scaling laws for these winds and show that they
can reproduce several observed properties of different sources. In
particular, the simple schematic diagram of a stratified accretion
disc wind shown in Fig. 5 is very similar to their MHD wind repre-
sented in their fig. 8. Magnetocentrifugally accelerated winds were
also previously studied by other authors, for instance Blandford &
Payne (1982) and Konigl & Kartje (1994), the latter suggesting a
similar unification scheme as Kazanas et al. (2012).
When observed with a certain inclination, the line-of-sight in-
tercepts distinct components of this stratified MHD accretion disc
wind, each with different velocity, column density, ionization and
projected distance. This model predicts that the inner part of the
flow should be faster and more ionized, being launched closer to
the black hole. The fact that the wind is launched from different
parts of the disc can provide an explanation of the large range in ob-
served blueshifted velocities, ionization and column densities of the
X-ray absorbers. In this picture, the UFOs and WAs would represent
the inner and outer parts of this accretion disc wind, respectively.
This is consistent with the relations shown in Figs 1-3 and with the
stratified accretion disc wind diagram shown in Fig. 5.
These self- similar solutions provide the simplest means of de-
riving a reasonable MHD wind model and allow us to derive scale
relations among different parameters to directly compare with ob-
servations. A realistic case is definitely more complicated than that
currently calculated [see Fukumura et al. (2010) and Kazanas et al.
(2012) for a detailed discussion of the limitations of these calcu-
lations]. In particular, these models consider the disc only as a
boundary condition; a fully self-consistent treatment should take
into account the accretion physics as well. However, they are so-
phisticated enough to provide at least a first order characterization of
these outflows and allow us to investigate the existence of underly-
ing relations/trends. Fukumura et al. (2010) focused their attention
on winds with a radial density profile of the type n(r) oc r -a , with
a = 1. This choice was driven by recent observational results on
the AMD of some X-ray WAs (Behar 2009). Instead, Blandford &
Payne (1982) adopted a slightly steeper density profile, a = 1.5. On
the other hand, Konigl & Kartje (1994) also considered solutions
with radial density slopes of a = 1 and 1.5, but the solution with
of — 1 was preferred because it was representing the ‘minimum
energy’ configuration. In Section 5.2 we estimated that the general
radial density profile of the observed outflows has a slope of roughly
a — 1.4, which is in between these two cases.
In particular, it is interesting to note the (qualitative) resemblance
of the scaling relations and dynamic ranges between the column
density, ionization, velocity and distance of the MHD wind model
of Fukumura et al. (2010) and Kazanas et al. (2012) (shown in the
left-hand panel of their figs 5 and 2, respectively) and our observed
relations in Fig. 1 (panels a and b) and Fig. 3(c). For instance, even
if they did not consider the case of the UFOs in their figure, we note
the wide range in ionization log£ ~ — 1 to 5 erg s _1 cm, velocity
v out ~ 10-10 000 km s -1 and distance log(r/r s ) ~ 4-10. They
also find the same trends of increasing ionization, column density
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies 13
and outflow velocity of the absorbers going from large to the small
distances to the central black hole.
Therefore, as already derived for the radiative case in Section
5.3.2, it is interesting to investigate the possible interpretation as
MHD winds using some more quantitative, although still somewhat
crude, considerations. The terminal speed of magnetocentrifugal
winds is proportional to the Keplerian velocity at the launching
region and it is a few times the escape velocity, u out = cov e sc where
co is a factor of the order of ~2-3 and it is the ratio between
the Alfven radius and the wind launching radius (Pudritz et al.
2007). The outflow velocity scales with the Keplerian speed as a
function of radius, so that the flow will have an onion-like layering
of velocities, the fastest inside and the slowest outside. As already
noted in Section 5.1, this velocity profile is directly consistent with
the one observed in Fig. 3(c) and reported in Table 3.
It is important to note that, since these winds are accelerated
by the action of magnetic torques from magnetic fields that are
embedded in the accretion disc, there is an intimate connection
between the mass-loss rate in the wind and the accretion rate on
to the black hole. In particular, one of the most profound scaling
relations for MHD winds is represented by the link between the
accretion and outflow rates, M acc ~ co 2 M out (Pudritz et al. 2007;
Reynolds 2012). Combining this with the accretion rate expressed
as M acc = L bo \/r]c 2 , where r] is the radiative efficiency, and the
definition of the momentum flux for the radiation and the wind we
can derive the relation:
^out,MHD — X T^rad? (9)
CO A Yj
where ft = v out /c, meaning that for a given velocity there is a
proportionality between the wind and radiation momentum rates
also for the MHD case. From equation (9) and the linear relation
Pout ~ P rad shown in Fig. 4 and discussed in Section 5.3 we find
that the radiative efficiency should be a few per cent for a typical
velocity of the UFOs of ft ~ 0. 1 in order to explain this proportion-
ality. Combining equation (9) with the expression for the outflow
momentum rate (5) and the definition of the ionization parameter
we can derive a relation between the velocity and the ionization
parameter for the MHD case as:
Unit, MHD — ~ 2 ( TTrl £’ (10)
47T m v C A \ T]CO z C f J
therefore, in this case we would roughly expect a direct proportion-
ality between the velocity and ionization u 0Ut oc %/rj. As already
done for the radiation pressure case, it is important to check the
general conditions under which an MHD wind can actually form,
imposing that the outflow velocity from equation (10) is equal or
higher than the escape velocity (Section 5.1). Using the approxima-
tion £ ~ L ion /A H r, the definition of the Eddington luminosity and
the Schwarzschild radius, we obtain
A > 2Vfr e (r)^ 2 C f , (11)
where k = L bo i/L E dd is the Eddington ratio, r e (r) is the Compton
scattering optical depth at the radius r = r/r s . Solving this equation
for r] and considering an average k ~ 0.15, an optical depth at the
radius logr ~ 1 of r e (r) ~ 0.1-1 (see Section 5.3.1 and Table
3) and the typical values of co ~ 2-3 and C f ~ 0.5, we obtain an
estimate of the radiative efficiency of the accretion disc of the Seyfert
galaxies considered here in order to be able to efficiently generate
MHD winds: r] < k/2\f?T e (r)co 2 Cf < 0.05-0.1. This value of the
radiative efficiency is comparable to the typical one derived for
quasars of r] ~ 0.1 (Soltan 1982; Elvis, Risaliti & Zamorani 2002;
Davis & Laor 2011). Regarding Seyfert galaxies in particular, some
authors suggested that the radiative efficiency for some of them
could be slightly lower than that (Bian & Zhao 2003b; Panessa
et al. 2006; Singh, Shastri & Risaliti 2011).
Therefore, from these considerations on the wind energetics and
the consistency with the expected overall structure/geometry and
velocity profile we derive that the observed outflows could be ef-
fectively accelerated by MHD processes.
However, it is important to note that this is only a partial conclu-
sion because a complete characterization of these outflows would
require the combined treatment of both radiation and MHD effects,
both important in AGNs. Some attempts in this direction have been
reported in the literature (Proga 2000, 2003; Everett 2005; Ohsuga
et al. 2009; Reynolds 2012). If these processes are acting simul-
taneously, we could naively expect to observe changes between
these two regimes. For instance, from equations (7) and (10) we can
speculate that the velocity should roughly show a proportionality
to f 1/2 from the radiation pressure term and to § from the MHD
part. It is tempting to compare this with Fig. 1(b), where we can
see that the residuals of the data with respect to the linear fit are
possibly consistent with a similar change of slope between these
two regimes going from lower to higher ionization. In fact, the lin-
ear regression fit is consistent with an intermediate slope of ~0.65.
Very close to the black hole and for higher ionization/velocities,
the MHD regime should always be the dominant part and it might
actually enter a regime that is eventually responsible for the ac-
celeration of jets. Further away from the black hole and for lower
ionization/velocities, radiation forces may contribute more to the
properties and dynamics of the flow, depending on the actual state
of the material and also the local disc luminosity.
5.4 Energetics and feedback impact
Having established that the correlations among the different outflow
parameters suggest an interpretation as a stratified, large-scale wind
with most probably both radiation and MHD mechanisms having a
role in the acceleration, then the next step is to, at least qualitatively,
investigate the contribution of these outflows on the expected AGN
feedback.
An important parameter to quantify the effective feedback contri-
bution of winds/outflows in bright AGNs is the ratio between their
mechanical power and the bolometric luminosity. Extensive numer-
ical simulations demonstrated that this value should be around ~5
per cent and can be as low as ~0.5 per cent (e.g. Di Matteo, Springel
& Hernquist 2005; Hopkins & Elvis 2010). Gaspari et al. (201 la, b,
2012b) recently demonstrated that outflows with these powers are
indeed adequate to produce sufficient feedback to quench cooling
flows in elliptical galaxies and to significantly eject gas without
overheating the galaxy itself. Since Seyferts are generally hosted
in spiral galaxies, which typically have less dense environments
compared to ellipticals, then the threshold could be even lower
than ~0.1 per cent (Gaspari et al. 201 la, b, 2012b).
Simulations of AGN outflows with characteristics equivalent to
UFOs have also been independently demonstrated to be able to sig-
nificantly interact not only with the interstellar medium of the host
galaxy but possibly also with the intergalactic medium. They can
provide a significant contribution to the quenching of cooling flows
and the inflation of bubbles/cavities in the intergalactic medium in
both galaxy clusters (e.g. Sternberg, Pizzolato & Soker 2007; Gas-
pari et al. 2011a) and especially groups (e.g. Gaspari et al. 2011b).
In Fig. 6 we plot the kinetic power of the outflows with respect
to the bolometric luminosity. We note again that the distribution of
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
14
F. Tombesi et al.
logo™)
Figure 6. Outflow kinetic power with respect to the bolometric luminosity.
The points correspond to the WAs (red open circles), non-UFOs (green
filled triangles) and UFOs (blue filled circles), respectively. The error bars
indicate the upper and lower limits and the points are the average between the
two. The transverse lines indicate the ratios between the outflow mechanical
power and bolometric luminosity of 100 per cent (solid), 5 per cent (dashed)
and 0.5 per cent (dotted), respectively.
points seems rather continuous from the WAs to the UFOs, the latter
having a higher power for a given source luminosity. As discussed
in Paper III and as evident from Fig. 6, the UFOs have indeed a
mechanical power clearly higher than 0.5 per cent of the bolometric
luminosity, with the majority of the values around ~5 per cent.
However, we should note that, as recently reported by Crenshaw &
Kraemer (2012), some of the WAs might actually reach the ~0.5
per cent level as well when their components are co- added together.
Therefore, these outflows, and in particular the UFOs, are clearly the
most promising candidates to significantly contribute to the AGN
feedback besides radio jets (e.g. Fabian 2012).
Theoretically, feedback from AGN outflows has been demon-
strated to clearly influence the bulge star formation and Supermas-
sive Black Hole (SMBH) growth and possibly also to contribute
to the establishment of the observed SMBH-host galaxy relations,
such as the Mbh-^ (e.g. King 2010a; Ostriker et al. 2010; Power
et al. 2011; Faucher-Giguere & Quataert 2012; Zubovas & King
2012). Similar and possibly even more massive and/or energetic
outflows might have influenced also the formation of structures and
galaxy evolution through feedback at higher redshifts, close to the
peak of the quasar activity at z ~ 2 (Silk & Rees 1998; Scannapieco
& Oh 2004; Hopkins et al. 2006).
From Fig. 3(a) we can note that on ~kpc scales the outflow is
very lowly ionized (log£ < 0 erg s -1 cm) and it could represent
the possible conjunction point with large-scale neutral/molecular
outflows recently found in some sources in other wavebands (e.g.
Nesvadba et al. 2008; Sturm et al. 201 1). Several models have been
suggested in order to explain their origin, but essentially all of them
rely on a two-step process in which an initial ~sub-pc scale fast
(v out > 1000 km s -1 ) AGN accretion disc outflow perturbs/shocks
the interstellar medium, sweeping it up on its way and then decel-
erates/cools (e.g. King 2010a; Faucher-Giguere & Quataert 2012;
Zubovas & King 2012). Therefore, it is tempting to check for possi-
ble evidence of these effects in our correlation plots. Their intensity
should be more prominent at large distances from the black hole.
Besides the large uncertainties, in Fig. 3(d) we can note a slight in-
crease of the mass outflow rate at > 100 pc (> 10 8 r s ) scales, possibly
suggesting some entrainment of surrounding material by the wind.
In Figs 3(e) and (f) we can also see a slight increase of the wind
momentum rate and mechanical power at those locations. These
evidence are roughly consistent with the relations reported in fig. 4
of Faucher-Giguere & Quataert (2012), who performed a detailed
study of the interaction of AGN winds with the surrounding environ-
ment and the different regimes of momentum/energy conservation
as the resulting shocked material propagates to large distances.
Observationally, we note that evidence for AGN feedback activity
driven by outflows/jets improved significantly in recent years but
there are still significant uncertainties, especially regarding the link
between the observed phenomenologies at small (~pc) and large
(~kpc) scales. Promising results on this line have been recently
reported for a few Seyfert galaxies, with the detection of bubbles,
shocks and jet/cloud interaction, some being also part of our sample
(e.g. Wang et al. 2010; Pounds & Vaughan 2011).
6 CONCLUSIONS
In order to investigate the possible relations between the UFOs,
mainly detected in the Fe K band through Fe xxv/xxvi absorption
lines, and the soft X-ray WAs we performed a literature search for
papers reporting the analysis of the WAs in the 35 type 1 Seyferts
of the sample defined in Paper I. The main results of our study are
as follows.
(i) The fraction of sources with reported WAs is >60 per cent,
consistent with previous studies. The fraction of sources with UFOs
is >34 per cent, >67 per cent of which showing also WAs.
(ii) We reported the main observed WA parameters, such as ion-
ization, column density and outflow velocity. Then, from these val-
ues, we estimated also the mass outflow rate, momentum rate and
mechanical power.
(iii) The large dynamic range obtained when considering all the
parameters of these absorbers together allows us, for the first time, to
estimate significant correlations among them. We find that the closer
the absorber to the black hole, the higher the ionization, column
density, velocity and therefore the mechanical power. In particular,
in the innermost part of the flow, at distances of log (r/r s ) ~ 1, we
find that the material can be mildly Compton-thick, N u ~ 10 24 cm -2 ,
highly ionized, log§ ~ 5 erg s -1 cm, and the velocity can approach
significant fractions of the speed of light.
(iv) In all the tests the absorber parameters uniformly cover the
whole parameter space, with the WAs and UFOs lying always at
the two ends of the distribution. This strongly indicates that these
absorbers, sometimes considered of different type, could actually be
unified in a single, large-scale stratified outflow observed at different
locations along the line of sight. The UFOs are likely launched from
the inner accretion disc and the WAs at larger distances, such as the
outer disc and/or torus. See Fig. 5 for a simple schematic diagram
of such a stratified wind.
(v) Given the high ionization and velocity of the outflows, and
a linear relation between the outflow and radiation momentum
rates, we argue that the only two viable acceleration mechanisms
are radiation pressure through Compton scattering and MHD pro-
cesses, the latter playing a major role. In particular, the overall
structure/geometry is more consistent with a stratified MHD wind
scenario.
(vi) Finally, as already discussed in Paper III, here we confirm
that these outflows, with the UFOs representing the most energetic
part, have a sufficiently high mechanical power (>0.5 per cent of
the bolometric luminosity) to provide a significant contribution to
AGN feedback. In this regard, we find possible evidence for the
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Unifying X-ray winds in Seyfert galaxies 15
interaction of the AGN wind with the surrounding environment on
large scales from the correlation plots.
In the future, in order to better quantify the predominance of
radiation pressure or MHD driving, we should extend the sample to
sources with lower and higher Eddington ratios. One would naively
expect outflows in lower luminosity AGNs to be dominated by
MHD processes while for bright quasars at the other end of the
distribution probably radiation pressure is more important. Seyfert
galaxies might represent an intermediate case between these two,
also suggested by their Eddington ratio of ~0.1. In this respect,
we plan also to directly test these hypotheses fitting the data with
detailed radiation and MHD wind models (e.g. Fukumura et al.
2010; Sim et al. 2010).
From an evolutionary point of view, it will be interesting to
compare the characteristics of similar outflows found in higher
redshift quasars in X-rays (Chartas et al. 2002, 2003, 2009; Giustini
et al. 2011; Lanzuisi et al. 2012) and also powerful, large-scale
outflows detected in other wavebands (e.g. Nesvadba et al. 2008;
Sturm et al. 2011).
Finally, we anticipate that the unprecedented high energy reso-
lution and sensitivity in the wide E ~ 0.1-10 keV energy band of
the microcalorimeter on board the upcoming Astro-H mission will
provide important improvements in this field, allowing for the first
time to simultaneously study in detail the absorbers in a wide range
of ionization states, column densities and velocities. We also note
that the unprecedented effective area of about 10 m 2 at 8 keV with
even moderate energy resolution of ~250 eV, such as the one pro-
posed by the Large Observatory for X-ray Timing mission to the
ESA Cosmic Vision, would allow us to detect UFOs in bright local
AGNs with high significance and at velocities up to ~0.7c, thanks
to the extended energy bandpass at higher energies.
ACKNOWLEDGMENTS
The authors thank the anonymous referee for the positive and con-
structive comments. FT thanks D. Kazanas, K. Fukumura, R. F.
Mushotzky for the useful discussions. MC acknowledges financial
support from ASI (contract ASI/INAF/I/009/10/0) and INAF (con-
tract PRIN-INAF-2011). RN was supported by an appointment to
the NASA Postdoctoral Program at Goddard Space Flight Cen-
ter, administered by Oak Ridge Associated Universities through a
contract with NASA. This research made use of the StatCodes sta-
tistical software hosted by Penn State’s Center for Astrostatistics.
This research has made use of data obtained from the High En-
ergy Astrophysics Science Archive Research Center (HE AS ARC),
provided by NASA’s Goddard Space Flight Center. This research
has made use of the NASA/IPAC Extragalactic Database (NED)
which is operated by the Jet Propulsion Laboratory, California In-
stitute of Technology, under contract with the National Aeronautics
and Space Administration. This research has made use of NASA’s
Astrophysics Data System.
REFERENCES
Abramowicz M. A., Fragile P. C., 2011, preprint (arXiv:l 104.5499)
Akritas M. G., Bershady M. A., 1996, ApJ, 470, 706
Akritas M. G., Siebert J., 1996, MNRAS, 278, 919
Ashton C. E., Page M. J., Branduardi-Raymont G., Blustin A. J., 2006,
MNRAS, 366, 521
Behar E., 2009, ApJ, 703, 1346
Bian W., Zhao Y., 2003a, MNRAS, 343, 164
Bian W. H., Zhao Y. H., 2003b, PASJ, 55, 599
Blandford R. D., Payne D. G., 1982, MNRAS, 199, 883
Blustin A. J. et al., 2003, A&A, 403, 481
Blustin A. J., Page M. J., Fuerst S. V., Branduardi-Raymont G., Ashton C.
E., 2005, A&A, 431, 111
Braito V. et al., 2007, ApJ, 670, 978
Cappi M. et al., 2009, A&A, 504, 401
Cardaci M. V., Santos-Lleo M., Hagele G. F., Krongold Y., Diaz A. I.,
Rodriguez-Pascual P., 2011, A&A, 530, A125
Chartas G., Brandt W. N., Gallagher S. C., Garmire G. P., 2002, ApJ, 579,
169
Chartas G., Brandt W. N., Gallagher S. C., 2003, ApJ, 595, 85
Chartas G., Saez C., Brandt W. N., Giustini M., Garmire G. P., 2009, ApJ,
706, 644
Costantini E. et al., 2007, A&A, 461, 121
Crenshaw D. M., Kraemer S. B., 2000, ApJ, 532, L101
Crenshaw D. M., Kraemer S. B., 2012, ApJ, 753, 75
Dadina M., Cappi M., Malaguti G., Ponti G., de Rosa A., 2005, A&A, 442,
461
Dadina M., Guainazzi M., Cappi M., Bianchi S., Vignali C., Malaguti G.,
Comastri A., 2010, A&A, 516, A9
Dauser T. et al., 2012, MNRAS, 422, 1914
Davis S. W., Laor A., 2011, ApJ, 728, 98
Di Matteo T., Springel V., Hernquist L., 2005, Nat, 433, 604
Elvis M., 2000, ApJ, 545, 63
Elvis M., Risaliti G., Zamorani G., 2002, ApJ, 565, L75
Emmanoulopoulos D., Papadakis I. E., McHardy I. M., Nicastro F., Bianchi
S. , Arevalo R, 2011, MNRAS, 415, 1895
Everett J. E., 2005, ApJ, 631, 689
Everett J. E., Ballantyne D. R., 2004, ApJ, 615, LI 3
Fabian A. C., 2012, ARA&A, 50, 455
Faucher-Giguere C.-A., Quataert E., 2012, MNRAS, 425, 605
Ferrarese L., Merritt D., 2000, ApJ, 539, L9
Fukumura K., Kazanas D., Contopoulos I., Behar E., 2010, ApJ, 715,
636
Gallo L. C., Miniutti G., Miller J. M., Brenneman L. W., Fabian A. C.,
Guainazzi M., Reynolds C. S., 2011, MNRAS, 411, 607
Gaspari M., Melioli C., Brighenti F., D’Ercole A., 2011a, MNRAS, 411,
349
Gaspari M., Brighenti F., D’Ercole A., Melioli C., 2011b, MNRAS, 415,
1549
Gaspari M., Ruszkowski M., Sharma R, 2012a, ApJ, 746, 94
Gaspari M., Brighenti F., Temi R, 2012b, MNRAS, 424, 190
George I. M., Turner T. J., Netzer H., Nandra K., Mushotzky R. F., Yaqoob
T. , 1998, ApJS, 114, 73
Ghisellini G., Haardt F„ Matt G., 2004, A&A, 413, 535
Giustini M. et al., 2011, A&A, 536, A49
Gofford J. et al., 2011, MNRAS, 414, 3307
Gofford J., Reeves J. N., Tombesi F., Braito V., Turner T. J., Miller L., Cappi
M., 2012, preprint (arXiv:1211.5810)
Gondoin R, Orr A., Lumb D., Santos-Lleo M., 2002, A&A, 388, 74
Haardt F., Maraschi L., 1991, ApJ, 380, L51
Halpern J. P„ 1984, ApJ, 281, 90
Holczer T., Behar E., Kaspi S., 2007, ApJ, 663, 799
Hopkins P. F., Elvis M., 2010, MNRAS, 401, 7
Hopkins P. F. et al., 2006, ApJS, 163, 1
Jimenez-Bailon E. et al., 2008, MNRAS, 391, 1359
Kaastra J. S., Mewe R., Liedahl D. A., Komossa S., Brinkman A. C., 2000,
A&A, 354, L83
Kaspi S. et al., 2000, ApJ, 535, L17
Kazanas D., Fukumura K., Behar E., Contopoulos I., Shrader C., 2012,
Astron. Rev., 7, 030000
King A. R., 2010a, MNRAS, 402, 1516
King A. R., 2010b, MNRAS, 408, L95
King A. R., Pounds K. A., 2003, MNRAS, 345, 657
King A. L. et al., 2012, ApJ, 746, L20
Konigl A., Kartje J. F., 1994, ApJ, 434, 446
Kraemer S. B. et al., 2005, ApJ, 633, 693
Krolik J. H., Kriss G. A., 2001, ApJ, 561, 684
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
16 F. Tombesi et al.
Krongold Y., Nicastro F., Elvis M., Brickhouse N., Binette L., Mathur S.,
Jimenez-Bailon E., 2007, ApJ, 659, 1022
Laha S., Dewangan G. C., Kembhavi A. K., 2011, ApJ, 734, 75
Lanzuisi G., Giustini M., Cappi M., Dadina M., Malaguti G., Vignali C.,
Chartas G., 2012, A&A, 544, A2
Lobban A. P., Reeves J. N., Miller L., Turner T. J., Braito V., Kraemer S. B.,
Crenshaw D. M., 2011, MNRAS, 414, 1965
Longinotti A. L., Bianchi S., Santos-Lleo M., Rodriguez-Pascual P.,
Guainazzi M., Cardaci M., Pollock A. M. T., 2007, A&A, 470, 73
Longinotti A. L. et al., 2010, A&A, 510, A92
Lusso E. et al., 2010, A&A, 512, A34
McKernan B., Yaqoob T., Reynolds C. S., 2007, MNRAS, 379, 1359
Markowitz A., 2009, ApJ, 698, 1740
Markowitz A., Reeves J. N., Braito V., 2006, ApJ, 646, 783
Markowitz A., Reeves J. N., George I. M., Braito V., Smith R., Vaughan S.,
Arevalo R, Tombesi F., 2009, ApJ, 691, 922
Matt G., Bianchi S., de Rosa A., Grandi R, Perola G. C., 2006, A&A, 445,
451
Nesvadba N. P. H., Lehnert M. D., De Breuck C., Gilbert A. M., van Breugel
W., 2008, A&A, 491, 407
Ohsuga K., Mineshige S., Mori M., Kato Y., 2009, PASJ, 61, L7
Onken C. A., Peterson B. M., Dietrich M., Robinson A., Salamanca I. M.,
2003, ApJ, 585, 121
Ostriker J. R, Choi E., Ciotti L., Novak G. S., Proga D., 2010, ApJ, 722, 642
Paggi A., Wang J., Fabbiano G., Elvis M., Karovska M., 2012, ApJ, 756, 39
Panessa F., Bassani L., Cappi M., Dadina M., Barcons X., Carrera F. J., Ho
L. C., Iwasawa K., 2006, A&A, 455, 173
Peterson B. M. et al., 2004, ApJ, 613, 682
Porquet D., Kaastra J. S., Page K. L., O’Brien P. T., Ward M. J., Dubau J.,
2004, A&A, 413, 913
Pounds K. A., Vaughan S., 2011, MNRAS, 413, 1251
Pounds K. A., Reeves J. N., King A. R., Page K. L., O’Brien P T., Turner
M. J. L., 2003, MNRAS, 345, 705
Pounds K. A., Reeves J. N., Page K. L., O’Brien P. T., 2004, ApJ, 616, 696
Power C., Zubovas K., Nayakshin S., King A. R„ 2011, MNRAS, 413, LI 10
Proga D., 2000, ApJ, 538, 684
Proga D., 2003, ApJ, 585, 406
Proga D., Kallman T. R., 2004, ApJ, 616, 688
Pudritz R. E., Ouyed R., Fendt C., Brandenburg A., 2007, Protostar. Planets
V, 951,277
Reeves J. N., Turner M. J. L., Pounds K. A., O’Brien P T., Boiler T., Ferrando
P, Kendziorra E., Vercellone S., 2001, A&A, 365, L134
Reeves J. N. et al., 2009, ApJ, 701, 493
Revnivtsev M., Sazonov S., Jahoda K., Gilfanov M., 2004, A&A, 418, 927
Reynolds C. S., 1997, MNRAS, 286, 513
Reynolds C. S., 2012, ApJ, 759, L15
Rozanska A., Czery B., Siemiginowska A., Dumont A.-M., Kawaguchi T.,
2004, ApJ, 600, 96
Sani E., Marconi A., Hunt L. K., Risaliti G., 2011, MNRAS, 413, 1479
Scannapieco E., Oh S. P, 2004, ApJ, 608, 62
Silk J., Rees M. J., 1998, A&A, 331, LI
Sim S. A., Miller L., Long K. S., Turner T. J., Reeves J. N., 2010, MNRAS,
404, 1369
Singh V., Shastri P, Risaliti G., 2011, A&A, 533, A128
Smith R. A. N., Page M. J., Branduardi-Raymont G., 2008, A&A, 490, 103
Soltan A., 1982, MNRAS, 200, 115
Starling R. L. C., Page M. J., Branduardi-Raymont G., Breeveld A. A., Soria
R., Wu K., 2005, MNRAS, 356, 727
Steenbrugge K. C., Kaastra J. S., Sako M., Branduardi-Raymond G., Behar
E., Paerels F. B. S., Blustin A. J., Kahn S. M., 2005, A&A, 432, 453
Sternberg A., Pizzolato F., Soker N., 2007, ApJ, 656, L5
Storchi-Bergmann T., McGregor P. J., Riffel R. A., Simoes Lopes R., Beck
T., Dopita M., 2009, MNRAS, 394, 1148
Sturm E. et al., 2011, ApJ, 733, L16
Svoboda J., Guainazzi M., Karas V., 2010, A&A, 512, A62
Tarter C. B., Tucker W. H., Salpeter E. E., 1969, ApJ, 156, 943
Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., Yaqoob T., Braito
V., Dadina M., 2010a, A&A, 521, A57 (Paper I)
Tombesi F., Sambruna R. M., Reeves J. N., Braito V., Ballo L., Gofford J.,
Cappi M., Mushotzky R. F., 2010b, ApJ, 719, 700
Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., Braito V., Dadina
M., 201 la, ApJ, 742, 44 (Paper II)
Tombesi F., Sambruna R. M., Reeves J. N., Reynolds C. S., Braito V., 201 lb,
MNRAS, 418, L89
Tombesi F., Cappi M., Reeves J. N., Braito V., 2012a, MNRAS, 422, LI
(Paper III)
Tombesi F., Sambruna R. M., Marscher A. P, Jorstad S. G., Reynolds C. S.,
Markowitz A., 2012b, MNRAS, 424, 754
Ulvestad J. S., Wilson A. S., 1984, ApJ, 285, 439
Ulvestad J. S., Antonucci R. R. J., Barvainis R., 2005, ApJ, 621, 123
Vasudevan R. V., Fabian A. C., 2009, MNRAS, 392, 1124
Vaughan S., Fabian A. C., Ballantyne D. R., De Rosa A., Piro L., Matt G.,
2004, MNRAS, 351, 193
Wandel A., Mushotzky R. F., 1986, ApJ, 306, L61
Wang J.-M., Zhang E.-P, 2007, ApJ, 660, 1072
Wang J., Fabbiano G., Risaliti G., Elvis M., Mundell C. G., Dumas G.,
Shinnerer E., Zezas A., 2010, ApJ, 719, L208
Wang J. et al., 2011a, ApJ, 742, 23
Wang J., Fabbiano G., Elvis M., Risaliti G., Mundell C. G., Karovska M.,
Zezas A., 2011b, ApJ, 736, 62
Winter L. M., 2010, ApJ, 725, L126
Wu X.-B., Han J. L., 2001, ApJ, 561, L59
Yaqoob T., McKernan B., Kraemer S. B., Crenshaw D. M., Gabel J. R.,
George I. M., Turner T. J., 2003, ApJ, 582, 105
Zhang S. N., Ji L., Marshall H. L., Longinotti A. L., Evans D., Gu Q. S.,
2011, MNRAS, 410, 2274
Zubovas K., King A., 2012, ApJ, 745, L34
This paper has been typeset from a TpX/LTpX file prepared by the author.
Downloaded from http://mnras.oxfordjournals.org/ at NASA Goddard Space Flight Ctr on June 18, 2013
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11